← Back to community benchmarks

Qwable-v1-oQ8-mtp

M4 Max (40c) · 128 GB · 8bit · 2026-06-21
User commentsymrex
Performance
195k
tokens
506.3
PP tok/s
42.6
TG tok/s
395019
TTFT (ms)
52.2
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.4.4
macOS macOS 26.3
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,321 94.8 35.4 GB view
4k 1,669 92.8 36.2 GB view
8k 1,674 90.7 36.5 GB view
16k 1,559 85.4 37.1 GB view
32k 1,339 79.4 38.5 GB view
64k 1,037 67.8 41.1 GB view
128k 689.6 53.0 46.5 GB view
195k 506.3 42.6 52.2 GB current
Batching Results
Batch Size TG tok/s Speedup
94.8 1.00×
140.2 1.48×
200.5 2.11×
257.6 2.72×